Please watch the video in full screen because the font size is a little small, also download the entire code from GitHub, the link is provided in the description so you can increase the font size and change the theme if you want in your local machine. Thank you for watching. Here is the code link: github.com/siddiquiamir/ML-MODEL-DEPLOYMENT-USING-FLASK

This is what I was looking for exactly

@StatsWire

2 жыл бұрын

Thank you

Thank you so much, it helped me to deploy in easy manner.

@StatsWire

Жыл бұрын

You're welcome!

Could you please clarify this moment: '''return render_template("index.html",prediction_text = "The flower species is {}".format(prediction))''' format(prediction) - displays the results as list, but I need to display the results as a table having all elements written from a new line /n. Is there some way to do it with format() function?

Thank you for a wonderful explanation!

@StatsWire

2 жыл бұрын

I am glad you liked it!

Appreciate the efforts, thank you❤

Very well explained. Good Job!

@StatsWire

Жыл бұрын

Glad you liked it

To the point video, thanks bro!

@StatsWire

2 жыл бұрын

You're welcome

Thank you, you save my assignment! 😂

@StatsWire

2 жыл бұрын

You're welcome :-)

Wonderful Explaination.

@StatsWire

2 жыл бұрын

! am glad you liked it! Thank you

This was very helpful thank you

@StatsWire

2 жыл бұрын

Glad it was helpful!

Informative 👍

@StatsWire

3 жыл бұрын

Thank you

Thank u sir.... Good Lecture and very helpful......

@StatsWire

Жыл бұрын

You're welcome!

Very nice tutorial

@StatsWire

3 жыл бұрын

Thank you

This is very help full for me

@StatsWire

Жыл бұрын

I'm glad you liked it. Thank you for your comment :-)

you are a life saver,,,thanks a lot. it was really helpful :)

@StatsWire

2 жыл бұрын

You're welcome:)

Thank a lot Sir!!!

@StatsWire

Жыл бұрын

You're welcome!

Dear Sir, Thank you very much for the video. It is really good. One small request, instead of printing the final decision, can you please provide details on how to return the probabilities for each class for given input values

@StatsWire

Жыл бұрын

HI, yes, we can print probabilites. You can use this function predict_proba

Please also elaborate on html file content n how are we migrating between different html pages

@StatsWire

Жыл бұрын

HTML is a different concept in itself that's why I did not go deeper into it but you can find online resources to understand the basics.

Thanks. Just one thing, I think you are missing the preprocessing for the prediction part.

@StatsWire

2 жыл бұрын

You're welcome. Yeah, that can be done. You can experiment with many more things.

Really learnt from this. But one thing sir, during the prediction stage, u only converted to float but you didn't use standardScaler on the prediction variables as you have done for the training set.Hope this won't affect the answer?

@StatsWire

2 жыл бұрын

Thank you, it would be good if you apply the same scaler object that you used to fit on the training data on unseen data you will get better results

@afeezlawal5167

2 жыл бұрын

Thanks for the quick reply. That is the problem here sir, How do I apply it sir? + let's say during the training, we did some feature extraction like onehotencoding, how can we apply it to the unseen data? Between, can you drop ur mail? THANKS.

@afeezlawal5167

2 жыл бұрын

@@StatsWire ., Please I need a reply to the above comment sir

@StatsWire

2 жыл бұрын

You'll have to store (i.e. pickle) your fitted LabelEncoders and OneHotEncoder. When you receive new data, you'll transform them via the already-fitted LabelEncoders and OneHotEncoder and then use your trained model to make the predictions. This way, the produced data will be in the exact format your models expects them to be

Thanks a lot

@StatsWire

2 жыл бұрын

You're welcome

No words only ❤

@StatsWire

Жыл бұрын

Thank you:)

Thank you so much Sir.

@StatsWire

2 жыл бұрын

You're welcome Susanth

@ssnayakmusic

2 жыл бұрын

Sir . I could create virtual environment using conda in the command prompt , but not able to integrate it with pPycharm. Could you please help me Sir ?

@StatsWire

2 жыл бұрын

@@ssnayakmusic I think you should not face any problem. Can you tell what errors are you getting? May be I can refer you some videos based on that

how could I access the website through any device and how do I make the data I plug in to be able to be used to retrain the model

Thanks 🙏

@StatsWire

2 жыл бұрын

You're welcome

Thanks for the tutorial. Can I use this if my model takes one image and several numerical features and then predicts? I mean how to handle the image part

@StatsWire

3 ай бұрын

I have not tried it with image, but ideally it should work with some minimal changes.

Nice video

@StatsWire

Жыл бұрын

Very nice

Thank you so much brother for the great explanation and amazing stuff. Best wishes to success you all work in all possible ways. 🤩 I was able to deploy it using VS Code.

@StatsWire

4 ай бұрын

Thank you for your kind words and lovely wishes!

Thanks. If I have categorical variables as X, how to handle them as predictors?

@StatsWire

Жыл бұрын

You need convert it to numerical first.

I am performing the same steps for my model. My site is getting created but when in give values to all the features and click predict, it gives me error saying that:- "ValueError: Specifying the columns using strings is only supported for pandas DataFrames Traceback (most recent call last)". I have tried giving both float and int values to the features on the site but get the same error. Please help me!

@StatsWire

2 жыл бұрын

Check if there are any missing values in the columns. Remove all missing values then follow all the steps you won't get this error

@StatsWire

2 жыл бұрын

For a better page layout use the jinja template and CSS. You will get it online.

Thanks, but I have a doubt, what if u have 2 models, and user can choose one of these model, give the input, how to do this, basically we will be having 2 pickle files

@StatsWire

2 жыл бұрын

Yes, you are correct we will have two pickle files. You can put them on two different ports. if they are connected to each other like taking the output of the model and giving it as an input to the other model then we can make only one final pickle file.

anaconda navigator also contain pycharm .. should we open that one ?? or should we have to install pycharm seperatly

@StatsWire

Жыл бұрын

I'm not sure but it has spyder I guess. We have to install pycharm separately.

thank you Sir, Is this method also applicable in the domain of deep learning ?

@StatsWire

4 ай бұрын

Yes, it is.

Hi sir, 1) please do let me know how to create a html of ml 2) this deployment is on local machine,please advise how to deploy on cloud like gcp,azure,aws.. 3) please also send any links related to other projects for practice. Thank you very much for your valuable information 😀....thanks & Regards

@StatsWire

2 жыл бұрын

Hi 1) To create more beautiful html of ml you have to use jinja. This is the link jinja.palletsprojects.com/en/3.0.x/ 2) You can deploy on cloud very easily, it's almost same you just need to create an account on heroku or aws. If you are comfortable with flask then you can deploy anywhere on cloud or azure, aws etc 3) You should also look at this video to know you can test on on postman. This is the video link: kzread.info/dash/bejne/eqyAz4-kis--l7g.html

@aa4734

2 жыл бұрын

@@StatsWire thank you vvv much sir, ...plz do an example with Google cloud platform it will be really appreciable.thanks and regards

@StatsWire

2 жыл бұрын

@@aa4734 Sure, I will make a video soon. Thank you for the great suggestion.

I am a newbie can u plzz tell me know what did actually request.forms.values() did in code on line no.17

@StatsWire

Жыл бұрын

It is getting the values in that line

can i use the same procedure for the model developed using jupyter notebook

@StatsWire

3 ай бұрын

Yes, it should work.

somewhere i read that there should be created a seperate environment for every project.. plz share link about the same (if you have already created the video on this issue) .

@StatsWire

Жыл бұрын

We can create a virtual environment also and then do the project. You can watch my video on pycaret installation where I've created a virtual environment.

@ddbshri

Жыл бұрын

@@StatsWire thank you

@StatsWire

Жыл бұрын

@@ddbshri You're welcome

I'm currently working on ML project on the topic of "Animal endangered species prediction"..but I'm don't know how to collect datasets for this So, can you sir please guide me in the dataset part please..?!!! Hope you will help me!! Waiting for ur reply

@StatsWire

2 жыл бұрын

Hi Harini, you can get the dataset online. You can find it on Kaggle or UCI UCI: archive.ics.uci.edu/ml/index.php Kaggle: www.kaggle.com

Hi Sir...Hi I'm deploying my first ML model using flask. I did exactly the same you explained in the video but when I'm trying to run app.py file using Spyder, it's saying "Restarting with windowsapi reloader" repetedely and not giving me the url. Please help

@FatimaSharif-bo8gr

27 күн бұрын

same problem i am facing, have you rectified yours please?

Hello I was able to do exactly as was done in the tutorial with same data and all but when I try to open the link created I get error saying "template not found" Please what can I do to resolve this

@StatsWire

Жыл бұрын

I guess you have not kept the HTML file in the correct location. Please check once.

VSCODE error while running app.py. DLL load failed while importing _multiarray_umath: The specified module could not be found.

@StatsWire

2 жыл бұрын

It is due to the conda path, you have to add Python.CondaPath in settings. Press Ctrl + Shift + P and select Terminal Configuration. Search for python.conda, and paste your conda path for example. C:\ProgramData\Anaconda3\Scripts\conda.exe

How did handle you the user input to be transformed to standard scaler ?

@StatsWire

2 жыл бұрын

Using standard scaler

@nirash8018

Жыл бұрын

@@StatsWire No shit Sherlock

Can You Please help me . Why am I getting this Error: ValueError: Number of features of the model must match the input. Model n_features is 132 and input n_features is 131

@StatsWire

2 жыл бұрын

Because there is mismatch between number of features in training data and prediction data. Make sure you have same number of columns in both train and prediction dataset.

@psgoat1111

2 жыл бұрын

@@StatsWire Thanks just now solved it out The issue was due to some reasons column name was inside trained model was different then name inside dataset

@StatsWire

2 жыл бұрын

@@psgoat1111 Great you have solved the error on your own

* Serving Flask app "__main__" (lazy loading) * Environment: production WARNING: This is a development server. Do not use it in a production deployment. Use a production WSGI server instead. * Debug mode: on * Restarting with watchdog (windowsapi) An exception has occurred, use %tb to see the full traceback. SystemExit: 1

@nishkarshnagle3363

Жыл бұрын

Maybe you are using it in jupyter file like .ipynb , try using same code in .py file

Good day bro.I have a model that i created using jupyter notebook .Which root should i take

@StatsWire

Жыл бұрын

Convert it to .py file and then you can follow all the steps.

I want to implement disease Prediction using symptoms In my Dataset I have 132 symptoms Can u suggest me something so that I can get 1 for selected symptoms and 0 for not selected symptoms

@StatsWire

2 жыл бұрын

You can convert symptoms to 1 and 0 and then train your model

@psgoat1111

2 жыл бұрын

@@StatsWire They are trained in 1 and 0 only But while talking inputs I am not able to understand how to do so

@StatsWire

2 жыл бұрын

@@psgoat1111 You have to convert your input data also into 1 and 0 then pass to the model

bro pls make a video for deployment with heroku

@StatsWire

2 жыл бұрын

Thank you for the suggestion. I will make it soon.

Got this error while loading the model Traceback (most recent call last): File "C:/ML Model Deployment/app.py", line 7, in model = pickle.load(open("model.pkl", "rb")) _pickle.UnpicklingError: invalid load key, 'A'. How to resolve. Please help

@StatsWire

2 жыл бұрын

I think you did not pickle it correctly. Can you please try the steps again one by one and see if it still throws an error

@srigayathriarunachalam3370

2 жыл бұрын

@@StatsWire tried multiple times sir still it throws the same error

@StatsWire

2 жыл бұрын

@@srigayathriarunachalam3370 This error is related to the code in app.py file. For this, we need to examine the code written inside the app.py file.

metadata-generation-failed. this error shows on installing sklearn . what to do?

@StatsWire

Жыл бұрын

Please paste the exact error.

Great video but I was straining my eyes to see your screen. Fonts were too small and it was really dark

@StatsWire

3 жыл бұрын

Sorry to hear that, I should have increased the font size I realized after reading your comment. Thanks for the constructive feedback, I will keep that in mind in future videos, and make good ones. I have posted the PyCharm code link please go and download it from GitHub, so you can see it on your system and change the font size as well as the dark theme to light one. Thank you again for the great feedback.

@davidobembe5302

3 жыл бұрын

Thanks a lot. It was not too bad though. I just had to put the video in full screen to be able to follow. Really informative video and clear explanations 👍

@StatsWire

3 жыл бұрын

@@davidobembe5302 Thank you for the feedback. I appreciate it.

I want to know what happened to the feature scaling you did in the model.py How the data that i give will be scaled??

@StatsWire

2 жыл бұрын

Those features will come on the same scale and it will give you stable results otherwise your results would not be stable if the features are on the different scales

@gracehim9154

2 жыл бұрын

@@StatsWire let's say i am doing nlp. Where my words are going through all processes like stop word removal lemmatization, tfidf etc. How will these be applied to unseen data that i pass???

@StatsWire

2 жыл бұрын

@@gracehim9154 You will have to pass your unseen data through all the steps you have done for the training data then only it will work. Like when you test your model that test data also goes through the same preprocessing steps along with the train data

I was getting this error "ValueError: X has 4 features, but DecisionTreeClassifier is expecting 3 features as input."

@StatsWire

2 жыл бұрын

Because after training the model using on 3 features you are passing 4 features for prediction. Please see if you are passing only 3 variables. Go through the code again it's a minor error you will be able to correct it.

Internal Server Error The server encountered an internal error and was unable to complete your request. Either the server is overloaded or there is an error in the application. This is what im getting after running on web. plz hepl me sir with this issue

@StatsWire

8 ай бұрын

Let me check

my ide is asking to register for the file type of model.pkl. how to resolve it?

@StatsWire

2 жыл бұрын

Hi, can you please paste the error? So I can help you

AttributeError: 'DecisionTreeClassifier' object has no attribute 'n_features_' I'm getting this error when I'm trying to run my app.py file. But thing is that I've saved the model that is RandomForestClassifier. So what is the error can u please help me. and thanks for your video. it makes easier to me to understand the whole code in Flask. Please help me with that error. thanks in advance.

@StatsWire

Жыл бұрын

Hello! DecisionTreeClassifier doesn't have such a method indeed.

Good video bro but my model has both string and float variables then how to perform flask code plz help me Thanks in advance

@StatsWire

3 жыл бұрын

No problem even my model has both string and float you just have to first convert to one-hot encoding in machine learning, and in flask, I have converted it into float so you will not face any problem. Just follow the steps. You can see it here in app.py "float_features = [float(x) for x in request.form.values()]" we are converting all features into float. Do not worry just follow the steps.

@umarfaruk-bq3qg

3 жыл бұрын

@@StatsWire can you please provide ur code if it is in github 😊

@StatsWire

3 жыл бұрын

@@umarfaruk-bq3qg Sure, this is the code link github.com/siddiquiamir/ML-MODEL-DEPLOYMENT-USING-FLASK

@umarfaruk-bq3qg

3 жыл бұрын

@@StatsWire In this are u use one hot encoding bro? By doing my app. py like this I get value error: could not convert string to float :'Year' Can u plz help me Thanks in advance

@StatsWire

3 жыл бұрын

First convert Year into int type in your model.py by typecasting for ex df["Year"] = int(df["Year"]) once you do that in app.py it will automatically convert this into float using the below function float_features = [float(x) for x in request.form.values()]

hi can u explain how we will deploy wordcloud model

@StatsWire

Жыл бұрын

Hello, you can create an object of that word cloud model, make an HTML file, and then deploy it. You will have to make changes in the HTML.

in which hosting website can we host

@StatsWire

3 ай бұрын

That depends where you want to host it. Mostly people host it locally.

where was the deployment section ?

@StatsWire

Жыл бұрын

It is in the same video

do we have to write something in "index.html"

@StatsWire

2 жыл бұрын

It has HTML codes already written. You can download it from my github repository and use it. But if you want to add something extra then you can edit the index.html file

@abhishekkumaragrawal1372

2 жыл бұрын

@@StatsWire thanks a lot

@StatsWire

2 жыл бұрын

@@abhishekkumaragrawal1372 you're welcome

how did u get that .csv file?

@StatsWire

Жыл бұрын

You can find code and csv file here: github.com/siddiquiamir/ML-MODEL-DEPLOYMENT-USING-FLASK

Can we use jupyter notebook

@StatsWire

2 ай бұрын

It will be difficult to maintain all the files

I have got the jinja2.exceptions.TemplateNotFound error

@StatsWire

2 жыл бұрын

Have you imported flask?

@manasiraut8091

2 ай бұрын

@@StatsWire even i have got the same error. i have imported flask

@StatsWire

2 ай бұрын

@@manasiraut8091 can you paste the exact error?

model.pkl is not created. why?

@StatsWire

3 ай бұрын

Can you please try it again from start. It should work.

why the web can't be reached sir?

@StatsWire

Жыл бұрын

Maybe some problem with your port. Please try again.

can you give that git repository link

@StatsWire

3 жыл бұрын

Yes, I have just updated the github link. Please go on this link to download the code. github.com/siddiquiamir/ML-MODEL-DEPLOYMENT-USING-FLASK

@krentwhite2668

3 жыл бұрын

Thanks buddy🤗

@StatsWire

3 жыл бұрын

@@krentwhite2668 You're welcome. A new video is coming today, how to test your model using postman after deploying on flask

@krentwhite2668

3 жыл бұрын

you are awesome dude! keep it up... we will support🔥🔥🔥

@StatsWire

3 жыл бұрын

@@krentwhite2668 Thank you

how to create html file sir

@StatsWire

Жыл бұрын

It's very easy. You can write code in any IDE like sublime text and just write the file extension as .html that's it

while running the app i getting following error

@StatsWire

Жыл бұрын

Which error?

@rangarajsingaravelu5594

Жыл бұрын

This is the development server don’t use the

@rangarajsingaravelu5594

Жыл бұрын

This is development server

@rangarajsingaravelu5594

Жыл бұрын

This is development server

@rangarajsingaravelu5594

Жыл бұрын

Don’t use the production server

Hello Sir... I find this video very educative while following the video and trying,i encountered an error in the app.py. the error says "Traceback (most recent call last): File "C:\ML Model Deployment\app.py", line 10, in model = pickle.load(open("model.pkl", "rb")) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ EOFError: Ran out of input " i wanted to see how your model.pkl file is but unfortunately i couldn't because it's not opening, and i noticed that my model.pkl file has nothing in it Sir if you can help me solve this problem i will really appreciate Thank you!

@StatsWire

27 күн бұрын

Can you check step by step where the problem is coming.

@FatimaSharif-bo8gr

27 күн бұрын

I did check Sir, but i can’t seem to find any problem